Login required to started new threads

Login required to post replies

Precision in coast down A/B aero testing
Quote | Reply
Hi all! Was wondering if anyone has a checklist on how to optimize quality of data collection when doing coast-down aero testing?

I have done a bit of Chung testing previously, but the only working U-shaped section of road I have found is about an hour away, whereas I have a straight 1km road with 5% elevation just a few minutes away. Would do simple A-B testing of helmets, bottle placement, mantis elevation% (5-10-15 etc), shoe/cleat combinations etc. Maybe along the road try different combinations of setups, but it quickly gets complex...

I would simply compare max speed at the end of the hill, running like AABBAB between two options. Absolute cda is not an issue. So I would need to control for these factors (changes for the duration of the test) what else? Thankful for any insight

Temperature (colder is better as aero factor more prevalent?)
Humidity
Total weight
Wind, strength/direction (looking for no-wind days)
Tire pressure (digital pressure meter)
Dry road
Speed/cadence sensor (not gps)
Repeatable position
Holding a straight line
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
scandinavianguy wrote:
Hi all! Was wondering if anyone has a checklist on how to optimize quality of data collection when doing coast-down aero testing?

I have done a bit of Chung testing previously, but the only working U-shaped section of road I have found is about an hour away, whereas I have a straight 1km road with 5% elevation just a few minutes away. Would do simple A-B testing of helmets, bottle placement, mantis elevation% (5-10-15 etc), shoe/cleat combinations etc. Maybe along the road try different combinations of setups, but it quickly gets complex...

I would simply compare max speed at the end of the hill, running like AABBAB between two options. Absolute cda is not an issue. So I would need to control for these factors (changes for the duration of the test) what else? Thankful for any insight

Temperature (colder is better as aero factor more prevalent?)
Humidity
Total weight
Wind, strength/direction (looking for no-wind days)
Tire pressure (digital pressure meter)
Dry road
Speed/cadence sensor (not gps)
Repeatable position
Holding a straight line

You can do precision testing with coast downs; it has one big advantage and several disadvantages.

The big advantage is that it doesn't require a power meter, or an accuate power meter. You can pretty much rely that when you're coasting your power is zero.

The disadvantages are

1) since you don't have power information, all of the usable information comes only from speed, so you need to have a wheel speed sensor -- you absolutely can't get good precision with only a speedometer based on GPS alone.

2) since you have to coast, only "half" your runs will count, so it takes much longer.

You don't want too steep of a hill, or you'll reach terminal velocity too quickly. The most useful information is in how quickly you accelerate (and decelerate), not the max speed.You'll get slightly better precision on a shallower hill. Rather than use max speed (aka "terminal velocity") as your metric, do (at least two) runs for each configuration at different starting speeds and use the speed data to reconstruct a VE profile.

Alternatively, you can use the hill for standard VE -- you don't need to find a "half-pipe." It's the reverse of the procedure above: climb the hill at two different speeds (two different powers). You can combine with the coast down the hill so that will double your runs and cut your testing time in half.

Both the "twice up the hill" and the "coastdown without a power meter" methods are mentioned in my white paper: see pages 98 and 108.
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
Well I would say you are better off not wasting your time trying. I have done the roll down testing for changes that make a big difference, let's say >10 watts, and been able to see noticeable differences, but the things you want to test are all likely to be smaller differences.

Finding a decent half pipe is way more difficult than most people realize, so I feel your pain. There is one solution hat might work for Chunging it without a half pipe. If you read through the slides that explain the Chung method, you will find the ""known" elevation solution for doing Crr and CdA simultaniously.. If you were to fix Crr and ride up the hill, Golden Cheetah will give you an virtical elevation profile for the hill. Then I think the only thing you need is to choose the CdA value that makes it so the total elevation you climb matches the total elevation gain of the VE profile. Do 4 or 5 runs up the hill and you should be able to get an idea of you CdA without a half pipe.

The weakness of this approach is that you need true elevation gain to get the CdA correct, sort of like you need true Crr to get CdA correct with a half pipe. This requires a survey of the end points, which most people don't have. You can, however, come up with an OK measure of elevation gain with your cell phone or bike computer. These don't provide great measurements of elevation gain, but as long as you commit to a single value, I would think you could get a decent relative CdA. The obvious downside of this method is all the editing you have to do of ride files because GC I and Aerolab aren't really set up to do this.

Just like on a half pipe, you probably want to vary the speed on some of your runs. The beauty of this method is that you get a great workout because you want to be hauling ass up the hill.
Quote Reply
Re: Precision in coast down A/B aero testing [RChung] [ In reply to ]
Quote | Reply
Excellent, many thanks for replying sir, really appreciate it! I will pull up your paper and have a good read before I get out there! Thanks for enlightening me about not looking at terminal velocity but acc/dec times. Is there a golden number for the appropriate incline for a ’shallow’ hill?

I would do the twice up the hill but my p1 pedals have been getting dropouts, so would drive me insane coming home to find the data useless!

Thanks again!
Quote Reply
Re: Precision in coast down A/B aero testing [grumpier.mike] [ In reply to ]
Quote | Reply
Hm, depressing if it is only useful for >10w differences. Sounds like I will have to do some more homework. Digging up true elevation numbers would be doable, there are really good survey services for the area in question!

Re hauling ass, exactly why I am attracted to this method haha, uphill could be intervals, then ’recovery’ while you switch setup!

Thanks!
Quote Reply
Re: Precision in coast down A/B aero testing [grumpier.mike] [ In reply to ]
Quote | Reply
grumpier.mike wrote:
Just like on a half pipe, you probably want to vary the speed on some of your runs. The beauty of this method is that you get a great workout because you want to be hauling ass up the hill.

You definitely want to vary the speed, both within and across runs. I do one run at about my max sustainable power for however long the run is (I have a hill I use that's just a bit more than a mile long at just under 5%, with VABMs (vertical angle benchmarks) at the start and end, so I know the exact change in elevation). Then I do a run at maybe half that power; then a run in-between. The total elevation gain is from the VABMs, so they have to match. I do try to follow the same line around each corner.
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
scandinavianguy wrote:
Hm, depressing if it is only useful for >10w differences.

I've done this method and gotten differences in CdA of .005 m^2, which is probably closer to 5w. The VE profile is a diagnostic, so you can tell whether the wind or some other interruption screwed up the estimate. That's helpful.
Quote Reply
Re: Precision in coast down A/B aero testing [RChung] [ In reply to ]
Quote | Reply
Isn’t there Cda variability within the pedal cycle?

Ie if you have a CDA of .200 then your actual CDA varies between .180 and .220 depending on if your feet are at 6/12 or 3/9.

I vaguely remember that from a study done on mannequins.

Maurice
Quote Reply
Re: Precision in coast down A/B aero testing [mauricemaher] [ In reply to ]
Quote | Reply
Ah, yes. I soft-pedal even when coasting. For me, it's easier than trying to remember exactly how to orient my legs and hold them still.
Quote Reply
Re: Precision in coast down A/B aero testing [RChung] [ In reply to ]
Quote | Reply
Ok, to vary power, that is good advice! And I will def soft pedal down, didnt think about that, was looking to keep pedals horizontal with same leg forward but easier to keep spinning! Thanks

What about outside temp, with the hour record they want a hot velodrome to lessen wind resistance, right? So would that mean that colder weather increases the proportion of wind drag (vs rolling resistance etc)?

Cheers
Quote Reply
Re: Precision in coast down A/B aero testing [mauricemaher] [ In reply to ]
Quote | Reply
Will update this thread once I have a few runs in the bank!
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
I try to remember to get measurements of baro pressure, temp, and humidity at the beginning and end of my test runs. (Humidity is less important than pressure and temp). I also know the elevation at the start. I use that information to calculate the air density, rho. If the temp or baro pressure changes a lot between the start and end of my runs, I'll recalculate rho though usually I don't have to. The exception is when I do my runs right at dawn; then, the temperature can be moderately different when I start and when I finish.
Quote Reply
Re: Precision in coast down A/B aero testing [RChung] [ In reply to ]
Quote | Reply
RChung wrote:
I try to remember to get measurements of baro pressure, temp, and humidity at the beginning and end of my test runs. (Humidity is less important than pressure and temp). I also know the elevation at the start. I use that information to calculate the air density, rho. If the temp or baro pressure changes a lot between the start and end of my runs, I'll recalculate rho though usually I don't have to. The exception is when I do my runs right at dawn; then, the temperature can be moderately different when I start and when I finish.

Do you use temperature to adjust Crr ?
Quote Reply
Re: Precision in coast down A/B aero testing [marcag] [ In reply to ]
Quote | Reply
Depends on what I'm testing for. If I need or want absolute accuracy, no, I don't. In those instances I use protocols that let me estimate both CdA and Crr.

But sometimes I'm just checking something informally, and in those cases if the temp is swings wildly then I'll take that into account.
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
I definitely wouldn't test in the cold just for higher air density. The temperature effect on drag is trivial compared to other things like speed. Ideally you want to test at temperatures you will be racing in. It reduces the error you incur from translating temperature effect on Crr. And another thing, if it's so cold you need warm clothes, then that data won't be applicable to warmer temperatures. If you are testing things besides your position though (like wheels maybe?) then it wouldn't matter.

Wind is what will kill you if you aren't measuring it (unless it's perfectly calm). Or traffic. Any things that change from one run to the next are problematic if they aren't accurately measured. If you can do A vs B and swap a lot that certainly helps, as random errors will be averaged out somewhat and drift should be noticeable.

I think you'd have better success doing VE. You don't need a half-pipe. Granted it's nice because you can do continuous laps and don't need to snip off the turning parts, but if you have one hill you can use for a turn you can use the CdACrr app with the fixed lap length and avoid snipping. There are other nice things about using the app too, like the ability to add a cheap airspeed meter, and getting numbers in the field.
Quote Reply
Re: Precision in coast down A/B aero testing [rruff] [ In reply to ]
Quote | Reply
Yeah, cold weather maybe not the best idea to test clothing, does shivering have an aero effect?

I am looking into getting a high precision anemometer, the cheaper ones seem to have 3-5% accuracy/error. Ideally maybe one that can log data, so changes between runs could be taken into account systematically. Suppose even small changes matter, like ’very little wind’ vs actually no wind. Thanks
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
If you use the CdACrr app, there is a particular anemometer that works with it. https://www.amazon.com/...eather/dp/B011WT29HO

I have it mounted ~18cm up and out from the basebar handle.
Quote Reply
Re: Precision in coast down A/B aero testing [rruff] [ In reply to ]
Quote | Reply
Thanks rruff, that is golden info, the meter looks really good! Have one om order now! Have you dome a writeup of your setup/protocol somewhere I missed? In the Platypus thread?

So you effectively use it as an aero stick? Or just to measure yaw? Thanks!
Quote Reply
Re: Precision in coast down A/B aero testing [scandinavianguy] [ In reply to ]
Quote | Reply
I don't have a writeup, but if you are using the app, this is the best place to discuss it: https://forum.slowtwitch.com/...oid_app__P5881854-7/

The app uses the anemometer for airspeed and temperature. It also measures barometric pressure but I think the app uses the phone's data instead. There is no yaw measurement.
Quote Reply